Impact of Guaranteed Income on Crime

 The Mincome Experiment (Manitoba Basic Annual Income Experiment) provides unique evidence on the impact of Guaranteed Annual Income (GAI) on crime and violence, demonstrating a significant negative relationship between the provision of unconditional cash payments and rates of crime. The study combining Mincome data with town-level crime statistics found that the guaranteed income affected both violent crime and total crime outcomes.

Mincome Context and Design

Mincome was an understudied GAI experiment conducted in the late 1970s, conceived as a central policy solution to poverty in Canada. While the primary attention of GAI experiments at the time was focused narrowly on labor market consequences and marital dissolution, the question of crime and violence received only one preliminary analysis. The new analysis focusing on Mincome aimed to expand the discussion to broader social consequences.

The study specifically focused on Dauphin, Manitoba, the Mincome "saturation" site where all residents were eligible for payments. This town-level experimental design was distinctive, unlike contemporary randomized controlled trials, as it allowed researchers to address macro-social questions concerning how treatment effects might spill over and affect community-level processes, such as aggregate crime and violence.

Mincome payments were equivalent to about $19,500 CAD (2014 dollars) for a family of four, and the scheme was structured so that working always increased a family's income. Importantly, the payment structure, which accounted for economies of scale, made "real the possibility of exiting bad or abusive relationships".

Impact on Crime Rates

Using a difference-in-difference regression comparing Dauphin's crime trends to those of similar control towns, the analysis revealed robust declines in crime.

1. Total Crime: Mincome was associated with a change amounting to 1,400 fewer total crimes per 100,000 people relative to similar town-years. Considering Dauphin’s overall crime rate averaged about 8,000 per 100,000 people, this change was considerable.
2. Property Crime: The experiment found a significant negative relationship between Mincome and property crime. This decline is consistent with the hypothesis that eliminating poverty reduces the straightforward economic incentives for property crime, aligning with rational choice theory. The results confirmed that Mincome leads to a reduction in property crime, meaning people commit fewer crimes when they have fewer reasons to do so. The analysis shows Mincome was associated with a substantial decline in the prevalence of this type of crime, amounting to between 726.1 and 959.7 fewer property crimes per 100,000 people (depending on the model used).
3. Other Crime: Mincome was also associated with a decline in "other crime" (a category including offenses like arson, disturbing the peace, and mischief), though this effect was much smaller and did not achieve statistical significance.

Impact on Violence and Mechanisms

The study found a strong, robust, and significant negative relationship between Mincome and violent crime rates.

• Magnitude of Effect: Mincome was associated with a change amounting to 350 fewer violent crimes per 100,000 people compared to other towns. This is a particularly large effect, given that Dauphin’s violent crime levels averaged about 600 per 100,000 people during the study period.
• Intimate Partner Violence as the Driver: While the link between economic variables and violent crime is often elusive, the strong effect on violence suggests that intimate partner violence is the most plausible candidate for the main kind of violence in decline. Statistics Canada noted at the time of Mincome that assaults constituted the largest portion of violent crime, and many of these offenses were domestic in nature, frequently arising from family disputes. Assaults made up 89% to 96% of violent crimes in Dauphin in the years where disaggregated data was available.

The sources speculate on two key mechanisms linking GAI payments to a decline in violence, both centered on mitigating financial hardship in relationships:

1. Reduction of Financial Stress: Exogenous income increases from Mincome could reduce financial stress and conflict in the family, which in turn reduces the likelihood of situational violence (the "income" effect). Economic hardship is strongly linked to higher levels of inter-partner violence.
2. Increased Bargaining Power and Exit Option: Mincome could change the balance of power within relationships, or increase a woman’s ability to exit from abusive marriages. Lack of independent economic resources is a significant reason women stay in violent relationships. Mincome provided the ability to start a single-person household, which, even as a threat, could improve bargaining power and reduce the chance of inter-partner assault.

The source material notes that future universal basic income policies, which allocate funds directly to individuals rather than households (like Mincome did), may be even more effective at reducing inter-partner violence because they are better positioned to offset familial power inequalities.

In the larger context of guaranteed income, the study concludes that Mincome demonstrates how providing a baseline level of financial stability serves to stabilize people’s everyday lives and temper conflict, addressing economic hardship and uncertainty—a key pathway to violent incidents.

The sources present strong key findings demonstrating a significant negative relationship between the Mincome Guaranteed Annual Income (GAI) experiment and rates of crime and violence in Dauphin, the saturation site. These findings were derived from a difference-in-difference regression analysis comparing Dauphin's crime trends to those of similar Prairie towns.

Negative Relationship with Crime and Violence

The analysis of town-level crime statistics during the Mincome experiment (1975–1977) revealed significant declines across major crime categories, supporting the hypothesis that unconditional cash payments reduce crime and violence:

1. Total Crime: Mincome was associated with a change amounting to 1,400 fewer total crimes per 100,000 people relative to similar town-years. Considering Dauphin's overall crime rate averaged about 8,000 per 100,000 people, this decline was substantial.
2. Violent Crime: The strongest and most robust finding was the significant negative relationship between Mincome and violent crime rates. Mincome was associated with approximately 350 fewer violent crimes per 100,000 people compared to control towns. This is a remarkably large effect, as Dauphin's average violent crime level was about 600 per 100,000 people during the study period.
3. Property Crime: The experiment found a significant negative relationship between Mincome and property crime. The results confirmed that Mincome was associated with a substantial decline in property crime prevalence, consistent with rational choice theory, which posits that reducing poverty decreases the economic incentive for such crimes. The specific change attributed to Mincome amounted to between 726.1 and 959.7 fewer property crimes per 100,000 people.
4. Other Crime: While Mincome was associated with a decline in "other crime" (a category including offenses like arson and disturbing the peace), the effect size was smaller and did not reach standard levels of statistical significance.

Context of Causal Mechanisms

The ability to link the decline in crime and violence directly to the Mincome payments strengthens the causal argument due to the experimental nature of the GAI, acting as an exogenous shock to people's incomes. The decline in crime supports the general hypothesis that income guarantees are inversely correlated with a range of crimes because, as material deprivation disappears, so should the associated social pathologies.

The sources extensively discuss the likely mechanisms driving the particularly strong decline in violent crime, linking it to intimate partner violence:

• Intimate Partner Violence as the Plausible Driver: At the time Mincome was conducted, Statistics Canada noted that assaults comprised the largest portion of violent crime, and many of these offenses were domestic in nature, stemming from family disputes. In Dauphin, assaults accounted for 89% to 96% of violent crimes during the period where disaggregated data was available. This evidence strongly suggests that intimate partner violence is the most plausible kind of violence to be impacted by Mincome.
• Reduced Financial Stress (Income Effect): The GAI provided exogenous income increases, which likely reduced financial stress and conflict within families, thereby reducing the likelihood of situational violence. Economic hardship is consistently linked to higher levels of inter-partner violence.
• Increased Bargaining Power/Exit Option: Mincome payments, which amounted to about $19,500 CAD (2014 dollars) for a family of four, provided the means for recipients to potentially start a single-person household. This option increased a woman's bargaining power within a relationship or provided a credible threat of exit, which could reduce inter-partner assault.

The sources conclude that the Mincome findings support the view that providing a baseline level of financial stability helps to stabilize people’s everyday lives and temper conflict, acting as an important policy tool to block off a key pathway to violent incidents.

The sources outline two general hypotheses regarding the impact of the Mincome Guaranteed Annual Income (GAI) on crime and violence, and they detail several specific mechanisms, particularly concerning violence, that explain the observed negative relationship.

General Hypotheses

The Mincome study was guided by two overarching hypotheses concerning the expected outcomes of providing unconditional cash payments:

1. (H1): Decline in Total Crime (Driven by Property Crime): Mincome was expected to be associated with a decline in total crime, driven primarily by a reduction in property crime. This hypothesis aligns with a rational choice perspective, which posits that a causal link exists between poverty and property crime because reducing poverty decreases the economic incentive (the benefit-to-cost ratio) of committing crimes. The logic is straightforward: a society that gives people fewer reasons for crime will see fewer crimes committed. A decline in other forms of crime often associated with poverty was also anticipated.
2. (H2): Decline in Violent Crime (Driven by Assaults): Mincome was also expected to be associated with a decline in violent crime. This reduction was anticipated to result from declines in both violent crimes linked to property crime and, critically, declines in assaults, which are largely composed of cases of intimate partner violence.

Mechanisms Linking GAI to Reduced Crime and Violence

The sources specify distinct mechanisms to explain the anticipated and observed declines in property crime and violent crime.

Mechanism for Property Crime

The mechanism linking Mincome to reduced property crime is seen as theoretically straightforward:

• Reduction of Economic Incentives: The GAI payments function as an exogenous shock to people's incomes. By effectively eliminating poverty, Mincome reduces the material deprivation and economic need that drive criminality, thereby reducing the straightforward economic incentives for property crime. This aligns with rational choice theory, which holds that people commit fewer crimes when they have fewer reasons to do so.

Mechanisms for Violent Crime (Focus on Intimate Partner Violence)

While the link between economic variables and violent crime is generally "more elusive" than the link to property crime, the sources focus on intimate partner violence (IPV) as the most plausible candidate for the strong decline in violence observed, as assaults (many of which are domestic in nature) constituted the vast majority of violent crimes in Dauphin.

The sources identify three primary mechanisms through which the guaranteed income could reduce inter-partner violence, all centered on mitigating financial stress and addressing economic dependency in relationships:

1. Reduction of Financial Stress ("Income" Effect): Exogenous increases in income from Mincome reduce financial stress and conflict within the family. Financial stress and conflict are known to heighten the risk of violent incidents, and reducing this stress is expected to improve relationship dynamics and reduce the likelihood of situational violence. This reduction in stress and conflict is often referred to as the "income" effect.
2. Increasing Ability to Exit Abusive Relationships: The GAI payments increase economic independence, offering women an option to exit from unhealthy or violent relationships through separation or divorce. The guarantee level, equivalent to about $19,500 CAD (2014 dollars) for a family of four, was structured to make "real the possibility of exiting bad or abusive relationships" by covering the costs of starting a single-person household.
3. Shifting Domestic Bargaining Power: Mincome reduces economic dependency, which in turn raises a woman's bargaining power within the partnership and makes threats to leave more credible. This shift in domestic power dynamics, even without an actual exit, can reduce the chance of inter-partner assault, particularly where violence is rooted in gender-based power inequalities.

The sources also note that original Mincome planning documents hypothesized that the GAI would:

• Reduce Juvenile Delinquency: By releasing family members (especially mothers) from work, the GAI would allow for greater parental control and attention, leading to a reduction in juvenile delinquency.
• Decrease General Community Crime and Mental Health Problems: By increasing the ability of the poor to participate in community organizations and enjoy a standard of living closer to the community average, crime and mental health problems were expected to decrease in poorer sections of the community.

Ultimately, the Mincome findings support the hypothesis that a reliable, automatic stream of cash income stabilizes people's everyday lives and tempers conflict, serving as an important policy tool to block off a key pathway to violent incidents.

The sources describe a rigorous methodology and specific data sets used to analyze the impact of the Mincome experiment on crime and violence, emphasizing the unique town-level design of the study.

Data Sources and Scope

The study combines two main sources of data to analyze the impact of Mincome:

1. Crime Statistics: The researchers obtained town-level crime data from Statistics Canada’s Uniform Crime Reporting surveys. These surveys provide information on different types of crime on a yearly basis for all Canadian towns and municipalities. The analysis uses data on crimes of violence and overall crime, spanning from the point of availability in 1972 to municipal boundary changes after 1980.
2. Socio-demographic Data: This crime data was merged with socio-demographic controls obtained from census data for each Prairie town, available for 1971, 1976, and 1981. Data for intervening years were linearly interpolated.

The analysis focused on Dauphin, Manitoba, the Mincome "saturation" site where all residents were eligible for payments, and compared it to all similarly-sized Prairie towns. Specifically, the control group included 15 Manitoba and Saskatchewan towns with mid-1970s populations between 5,000 and 50,000.

Dependent Variables and Measurement Challenges

The analysis included four main dependent variables (crime rates per 100,000 people):

1. Overall crime rate.
2. Violent crime rate.
3. Property crime rate (including breaking and entering, theft, and auto theft).
4. "Other" crime rate (including offenses like arson, disturbing the peace, and mischief).

The study notes that, ideally, further disaggregated data for violent crime would be used, but this was unavailable. However, the available data showed that assaults constituted the vast majority (89% to 96%) of violent crimes in Dauphin in the years where disaggregated data was available (1972–1973), making the violent crime rate a plausible proxy for changes in intimate partner violence.

The researchers acknowledge that they rely on reported crime statistics, which do not cover the full universe of crime, particularly due to the high prevalence of underreporting in intimate partner violence. However, since the analysis relies on change-scores (the difference-in-difference method), underreporting would only bias the estimates if the Mincome treatment itself affected the rate of underreporting, or if trends in underreporting did not linearly map onto crime prevalence trends. The researchers assume the latter sources of bias are minimized, allowing results to reasonably extrapolate to crime trends more generally.

Unique Experimental Design and Methodology

The Mincome study utilized a distinctive town-level experimental design in Dauphin, known as a "saturation" site, where every resident was eligible for the GAI. This design is critical to the methodology because it allows researchers to study macro-social questions, such as how treatment effects might spill over and affect community-level processes like aggregate crime and violence. This contrasts sharply with later GAI experiments, which were generally set up as Randomized Controlled Trials (RCTs) focusing on individual recipients, which make macro-social questions "unaskable by design".

The primary estimation method was a multi-year difference-in-difference (DiD) regression that includes town and year fixed effects.

• Difference-in-Difference (DiD): This method tests whether changes in crime rates in Dauphin (the treatment site) deviate from those of other similar Prairie towns during the Mincome experiment (1975–1977). This approach approximates the unobserved counterfactual (what Dauphin's crime rate would have been without Mincome) by using the change-over-time trends of the control towns.
• Fixed Effects: This technique leverages only within-unit (within-town) variation, thereby eliminating biases caused by stable, unobserved heterogeneity between towns (e.g., fixed cultural differences or baseline poverty levels, or idiosyncrasies in local policing). The inclusion of time-period fixed-effects also controls for time-varying heterogeneity that is uniform across all units (e.g., provincial crime or economic trends).

Socio-demographic Controls

To ensure that the observed crime declines were truly attributable to Mincome and not other simultaneous changes, the models included time-varying controls, proxying for population-level changes that commonly correlate with crime:

• Socioeconomic Status: Rate of labor-force participation, average family income, and percentage of high-school graduates.
• Age: Percentage of the population between 20 and 24 years old.
• Inter-partner Violence Proxies: Percentage of single-female households, percentage of divorced females, and the ratio of female-to-male average incomes (used as a proxy for domestic power asymmetry).

Robustness Checks

To address potential limitations in the data and methodology, the researchers conducted two sets of robustness checks:

1. Placebo Test: The researchers tested if the findings were driven by an unobservable time-varying trait unique to Dauphin. This was done by switching the "treatment" years to a non-experimental period (1978–1980). This test found no evidence for an effect of this Mincome placebo, suggesting the observed decline was tied to the experiment years.
2. Synthetic Control Method: This method was used to address the concern that the selected control towns might be an inadequate comparison group. It constructed a "synthetic control" for Dauphin using a weighted average of the best-matching control towns. Results from the synthetic control analyses confirmed the estimates obtained using the fixed-effects regressions for all outcome variables.

The virtue of using an experiment like Mincome, as noted by the sources, is the ability to link outcomes with a fundamentally exogenous cause—the availability of Mincome payments—thereby strengthening the case for making a causal argument about the impact of guaranteed income on crime.